MLIR in Action: A Practical Guide to Scalable Model Optimization and Hardware Acceleration With OpenXLA, IREE and Mojo

By: Orian, Reid

MLIR in Action: A Practical Guide to Scalable Model Optimization and Hardware Acceleration with OpenXLA, IREE, and Mojo

Unlock the full power of machine learning optimization and next-generation compiler design with MLIR in Action - your complete, hands-on guide to mastering the Multi-Level Intermediate Representation (MLIR) ecosystem.

Built for engineers, researchers, and AI practitioners, this book takes you on a step-by-step journey through the core concepts, workflows, and real-world applications of MLIR - the backbone of modern compiler infrastructures like OpenXLA, IREE, and Mojo. Learn how to optimize, transform, and deploy machine learning models efficiently across CPUs, GPUs, and custom accelerators using a unified and extensible compiler stack.

Inside this practical guide, you will discover how to:

- Understand the architecture and principles of MLIR in depth.

- Build, extend, and debug custom MLIR dialects and passes.

- Integrate MLIR with leading frameworks such as TensorFlow, PyTorch, and Mojo.

- Leverage OpenXLA and IREE for portable, high-performance model deployment.

-Automate builds, CI/CD pipelines, and cloud deployment for scalable production systems.

-Visualize, profile, and debug IR flows for performance tuning and optimization.

- Stay ahead of the curve with insights into emerging compiler standards, AI-driven optimization, and future MLIR trends.

From theory to practice, every chapter blends clear explanations with real code examples, text-based flowcharts, and implementation checklists. Whether you are optimizing large-scale AI workloads or exploring compiler-based acceleration, MLIR in Action gives you the tools to move from concept to production with confidence.

Perfect for:

- Machine Learning Engineers

- Compiler Developers

- AI Infrastructure Architects

- Systems Programmers

- Researchers exploring hardware-aware AI optimization

MLIR in Action bridges the gap between research and real-world deployment - helping you build the scalable, efficient, and future-ready AI systems that power the next wave of machine learning innovation.

• Format: Paperback
• Language: English
• ISBN: B0FY5YR4H7
• ISBN13: 9798272080059
• Release Date: October 2025
• Publisher: Independently Published
• Length: 246 Pages
• Weight: 1.28 lbs.
• Dimensions: 0.5" x 8.5" x 11.0"